The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): The skill follows security best practices by using environment variables (process.env.RESEND_API_KEY) instead of hardcoding credentials. No sensitive file access or unauthorized data exfiltration was detected.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): The code utilizes the official resend Node.js package. No suspicious remote scripts or piped execution patterns (e.g., curl|bash) were found.
[Indirect Prompt Injection] (LOW): The skill exposes a vulnerability surface where untrusted external data could influence agent behavior if this integration is used in an automated pipeline.
Ingestion points: The webhook handler in /api/webhooks/resend processes incoming email content (event.data.subject, event.data.html, event.data.text) which is attacker-controlled.
Boundary markers: Absent; there are no delimiters or instructions to ignore embedded prompts within the forwarded email content.
Capability inventory: The skill has the capability to send emails (resend.emails.send) and manage audience contacts (resend.contacts.create).
Sanitization: Absent; the content is directly interpolated into an HTML template for forwarding, which could be exploited if a downstream agent or user processes the forwarded email.

resend-integration